@eyedeekay
&eche|on
&kytv
&zzz
+R4SAS
+RN
+StormyCloud
+T3s|4
+acetone
+dr|z3d
+hagen
+hk
+lbt
+orignal
+postman
+segfault
+weko
+wodencafe
Arch
Danny
DeltaOreo
FreefallHeavens_
Irc2PGuest10722
Irc2PGuest16699
Irc2PGuest21882
Irc2PGuest27222
Irc2PGuest39482
Irc2PGuest59134
Irc2PGuest95432
Leopold
Nausicaa
Onn4l7h
Onn4|7h
Sisyphus
Sleepy
Soni
T3s|4_
Teeed
aeiou
aisle1
ardu
b3t4f4c3__
bak83_
boonst
bpb
cumlord
dickless
dr4wd3_
eyedeekay_bnc
poriori
profetikla
qend-irc2p
radakayot_
rapidash
shiver_
solidx66
thetia
u5657
uop23ip_
w8rabbit
x74a6
orignal
zzz, about streaming signatures
orignal
same idea as for datagrams
orignal
if SYN comes through a ratchets session we might omit signature
zzz
eyedeekay, good news, izpack 5.2.4 is out with the java 21+ fix. Please update and test your CI
zzz
I'll be using it for the release
zzz
eyedeekay, re: github.com/i2p/i2p.i2p/pull/100 - what is the issue you refer to that this fixes?
dr|z3d
we use alpine latest in +, no reported issues.
dr|z3d
(I'd far rather set and forget than have to keep updating version strings as new more suitable versions appear)
zzz
OP said 'various issues', eyedeekay said 'the issue', neither elaborated
dr|z3d
Don't you just love those :)
zzz
OP said 'various issues', eyedeekay said 'the issue', neither elaborated
zzz
eyedeekay, just looked at git.idk.i2p/i2p-hackers/i2p.firefox/-/issues , gave you a new one, and looks like you should go thru all the 2-year-old ones and close what you can
zzz
orignal, re: streaming, yes, we discussed it briefly on tuesday
eyedeekay
Sorry had errands to run this morning issue was related to outdated Java/libs in container, reports were in #i2p of the Docker container being unusable due to issues with Java 17 and mortbay. First noticed on TrueNAS then was observed on plain Docker. phil who has been helping with go-i2p recognized the issue and submitted the fix, I tested it.
zzz
ok eyedeekay thanks. Did you triple-check that "temerin" and the download URLs he changed are legit, and we aren't getting supply-chain-hacked?
eyedeekay
Yeah Termurin/the temurin container is legit it's what they changed the Eclipse OpenJDK distribution name to a few years ago
zzz
ok great
zzz
can you give me a short oneliner description of the issue for the upcoming release notes?
eyedeekay
Updated outdated docker container and container libraries
zzz
thx
eyedeekay
And also probably credit urgentquest from Github
zzz
don't usually give credit in roadmap or release notes but feel free to add it to history.txt
eyedeekay
Will do
zzz
I have my fingers in i2ptunnel today, seeing if I can generate a 5,4 LS2
zzz
lets see what explodes
zzz
03/06 12:33:33.527 CRIT [nal Reader 4] 2p.data.i2cp.I2CPMessageReader: Uncaught I2CP error
zzz
java.lang.IllegalArgumentException: Unsupported algorithm
zzz
that didn't take long
zzz
this is going to be messy
orignal
zzz, that's for datagram not streaming
orignal
for streaming I'm thinking to do it now
orignal
just not verify SYN signature if from ratchets session
zzz
we discussed streaming briefly also
orignal
well I didn't ask my question yet )))
zzz
<zzz>if we don't need 'from' or 'signature' any more, shouldn't we fix i2cp and streaming first?
orignal
my question is
orignal
is it possible to have stream on behalf of third party?
zzz
please explain more
orignal
someone sends SYN packet and LS that doesn't match ratchets static key
orignal
so, they send LS and From matching that LS
orignal
but it's from another destination, not the one for ratchets session
orignal
that's the key question for next steps
zzz
don't think so but would have to think about it more
orignal
if we require LS matching session than we can filter LS-es first
orignal
no unsolicited LS, but own
zzz
this sounds very messy and hard
orignal
another LS must be response to our lookup
orignal
why?
orignal
you got LS and check
orignal
1. Did we request it?
orignal
2. If we didn't, does it match sessions's static key
orignal
?
orignal
very simple
zzz
it's a big layering violation, smashing together two different protocol layers, and I have i2cp in between
orignal
how?
orignal
do we send peer-s LS though I2CP to client?
orignal
I donb't remember that we do
zzz
handling leasesets in ratchet
zzz
no
orignal
then why two direrent levels?
zzz
but as I said tuesday, there's no way to pass 'from' through i2cp
orignal
we are talking about LeaseSets now
zzz
yeah but if you strip the from and sig fields out of streaming then streaming doesn't know where it's from
orignal
* [zzz] (zzz@5fweozxzwtgzln7sljuistexwqoias4iz3csub3jlnopxkubjaga.b32.i2p): zzz realname
orignal
if you don't pass "From" through I2CP how this one works?
zzz
in streaming. I mean separately
orignal
no I don't want to strip From
orignal
only signature
orignal
because it's expesive
zzz
then stop verifying
orignal
that's what I'm going to do
orignal
but I also want to stop signing
zzz
I'm not recommending you stop verifying, but do what you want
orignal
also you don't have "From" at rachets level anyway
orignal
you have only static keys
orignal
e.g. it might be many "From"s matching the same static key
zzz
removing signing would be a big protocol change, I've added your idea to the list of proposals we're waiting for from you
zzz
yeah the shared static key thing makes it harder
zzz
aka subsessions
orignal
why? it's just a flag
zzz
but to not have huge PQ SYN packets you'd have to get rid of the From part also
orignal
but you always have it like EdSA/DSA shared dest
orignal
how long is PQ public key?
zzz
yeah and I'm just starting on shared Ed/PQ dests type 5,4
zzz
see proposal 169
orignal
btw, why can't we pass B32 in From field?
orignal
sice we are supposed to have peer's LS anyway?
zzz
jrandom didn't have b32
orignal
are you jrandom? me ethier
orignal
why we can't change it?
zzz
we can change anything we agree on, write a propsal
orignal
I will do
orignal
but mny question is maybe it was a reason for it
zzz
since I'm not jrandom, I can't answer
orignal
so you don't have any resons in mind
orignal
I will write a proposal and amendment for Datagram2
orignal
"use b32/b33 instaed full address"
orignal
well b32 only, no b33 ofc
zzz
no reason that I know of, before my time, but our git history goes back to April 2004 if you want to look around
zzz
I think you mean binary hash, not b32
orignal
sure 32-bytes SHA256
zzz
working thru 5,4 issues to get tunnel tests working
orignal
5,4?
orignal
I remeber you had an issue with one time keys
zzz
PQ + ECIES LS2
zzz
a lot of unchecked exceptions, making it pretty easy to find issues and fix them
zzz
think I got it kinda working
orignal
I will join on 11-th
orignal
opensl 3.5 alpha
zzz
yup
zzz
something's not right with tunnel builds
zzz
this is going to be a while
zzz
baby steps